Next Generation Sequencing of Hematologic Neoplasms
Todd W. Kelley, M.D.Associate Professor of Pathology
University of UtahMedical Director of Molecular Hematopathology
ARUP LaboratoriesSalt Lake City, Utah USA
1. List a few examples of the types of NGS tests
2. Describe the clinical utility of NGS technology in the context of testing of hematologic neoplasms
Learning Objectives:
Outline• NGS background• Overview of types of clinical NGS tests• NGS panels• Single gene tests
– Lymphoid clonality testing by NGS– BCR-ABL1 kinase domain sequencing
• Conclusions
Next Generation Sequencing (NGS) Impact of NGS
1st generation sequencing - Sanger sequencing– utilizes chain terminating dideoxynucleotides– slow and laborious, method has been relatively unchanged for ~30
years– data = mixture of sequences– sequence data can be reviewed manually– poor sensitivity for detection of variants (~15-20%) – relatively long contiguous sequence can be generated (>600bp)
NGS - also known as massively parallel sequencing
– parallel single molecule sequencing– millions of small fragments of DNA are immobilized on a solid surface,
amplified (copied), and sequenced simultaneously– during sequencing a signal (light, pH change) is detected when a base
is incorporated– short contiguous sequences (reads) are generated – reads are aligned to a reference sequence and analyzed – analysis is computationally intense
Comparison of NGS applications
The power of NGS
‐Study performed by the Cancer Genome Atlas Research Network
‐200 cases of de novo adult AML subjected to whole genome (50) or whole exome (15) sequencing
‐Tier 1 – coding changes or splice sites
‐average of 13 overall (all tiers) mutations per case
‐23 genes significantly mutated (>5% of cases)
‐majority of cases demonstrated more than 1 clone basedon distinct clusters of variant allele frequencies (VAFs)
Papaemmanuil E et al. Blood 2013;122:3616-3627©2013 by American Society of Hematology
Higher overall number of oncogenic mutationscorrelated with worse outcome
‐738 patients with MDS, MDS‐MPN‐111 cancer associated genes were sequenced by NGS (gene panel)‐78% of patients had 1 or moreoncogenic mutations‐No systematic differences betweenDNA derived from bone marrow orperipheral blood
Clinical impact of somatic mutations
• SF3B1 mutations are associated with favorable outcome
Clinically important information is derived from large scale genetic analysis by NGS:
The example of MDS
Malcovati L et al. Blood 2014;124:1513-1521©2014 by American Society of Hematology
308 pts w/ myeloid neoplasmsMDS: 245MDS/MPN: 34AML‐MDS: 29
111 gene mutation panel
*Almost all patients with RARS (refractory anemia with ring sideroblasts) had an SF3B1mutation
Clinical applications of NGS in hematology• Clinical applications:
– Whole genome sequencing (entire genome ‐ ~3B base pairs)– Whole exome sequencing (~30M base pairs)
• Sequencing limited to protein coding regions representing ~1% of genome
– Mutation panels• Myeloid
• AML prognostic markers – FLT3, NPM1, CEBPA, ASXL1, IDH1/2• Myelodysplastic syndromes (MDS) – cohesin and spliceosome genes frequently mutated• Myeloproliferative neoplasms (MPNs) – JAK2, CALR, MPL, ASXL1• Pan myeloid panels
• Lymphoblastic leukemia and mature lymphoid neoplasms• Ph‐like lymphoblastic leukemia• Diffuse large B cell lymphoma (BCR pathway mutations)• Mutations associated with T cell lymphoproliferative disorders (JAK‐STAT pathway mutations)• Pan lymphoid panels
• Congenital disorders – bone marrow failure syndromes, congenital hemolytic anemias
– Detection of complex genomic abnormalities ‐ copy number variants (CNVs) and translocations
– Analysis of single genes with high complexity • Ex. lymphoid clonality and IGH or TRG/TRB genes
Whole genome sequencing• Many of the biomarkers we now know to be
important were discovered in whole genome sequencing studies (ie. DNMT3A, IDH1/2, etc)
• Not routinely performed in the clinical lab – Would need paired normal tissue for tumors– Time consuming– Expensive– Yields relatively low coverage (~30X) so results may
be difficult to interpret, especially with low tumor burden
• Benefit: Not limited to selected targets
Spectrum of mutations in myeloid malignancies AML, MDS, MPN and MDS/MPN overlap disorders
FLT3KITJAK2MPLKRAS/NRASPTPN11NF1CSF3R
Cell signalingCEBPARUNX1GATA1/GATA2PHF6ETV6
Transcription
DNMT3ATET2IDH1/IDH2
ASXL1EZH2SUZ12KDM6A
Epigenetics
SF3B1SRSF2ZRSR2U2AF1
Splicing
STAG2SMC1ASMC3RAD21
Cohesin complex
TP53NPM1
Cell cycle
Myeloid malignancies
Matynia et al et al. 2015. Archives of Pathology and Laboratory Medicine.
There is often a complex subclonal architecture in myeloid malignancies
Matynia et al et al. 2015. Archives of Pathology and Laboratory Medicine.
Pre diagnosis Diagnosis RelapseEx. clonal hematopoiesis of uncertain significance (CHIP)
Variant Associations
From: Tietz textbook of Clinical Chemistry and Molecular Diagnostics, 6th Edition
• Tiered strategy– A variety of systems are in use and this
area currently lacks a uniform standard
Mutation panels: Variant reporting
Higher tiers – more likely to be pathogenic or actionable
Lower tiers – less likely to be pathogenic or likely or known germline polymorphism
NRAS c.37G>C, p.Gly13Arg
TET2 c.5284A>G, p.Ile1762Val
Variants of unknown significance (VUSs)
• 52 year-old female presented with easy bruising and fatigue– CBC: WBC – 33 K/uL, Hgb – 9.6
g/dL, Platelets – 12,000 K/uL– Flow cytometry on BM aspirate:
large CD34 negative atypical myeloid blast population (48% of leukocytes)
– BM morphology – Acute myeloid leukemia
– Cytogenetics/FISH – normal karyotype
Clinical Scenario #1
Clinical scenario #1 -mutationsMutation panel testing by NGS:
Tier 1 variants:1. NPM1 c.860_863dup, p.Trp288fs-Variant frequency 35.5%-Associated with good prognosis except when a FLT3-internal tandem duplication mutation is present.
2. FLT3 c.1802_1803ins45, p.Leu601_Lys602ins15-Variant frequency 30.0%-Associated with early relapse and poor overall survival.
3. DNMT3A c. 2645G>A, p.Arg882His-Variant frequency 41.2%-Commonly seen with NPM1 mutations in patients with CN-AML-DNMT3A R882 mutations are associated with poor outcome when compared to NPM1 mutated AML patients without DNMT3A mutations
Conclusion – Poor prognosis; patient should proceed to BM transplant
Clinical scenario #2
• 75 y/o male with complaint of fatigue and history of primary myelofibrosis
• CBC: – WBC: 40.05 k/uL– Hgb: 14.9 g/dL– MCV: 76.5 fL– Plts: 205 k/uL
• Cytogenetics: 46, XY, inv(12)
Clinical scenario #3
• Ph-like lymphoblastic leukemia– Gene expression profile similar to Ph+
(BCR-ABL1+) lymphoblastic leukemia but do not have t(9;22);BCR-ABL1.
– Affects children (10% with standard risk ALL) and adults (~20%).
– Variety of molecular abnormalities that activate tyrosine kinase signaling pathways including rearrangements (CRLF2, ABL1, ABL2, etc) as well as mutations involving FLT3, IL7R, SH2B3, etc)
– Worse outcome compared to non-Ph-like ALL
– Benefit from tyrosine kinase inhibitor therapy
– A properly designed NGS panel can assess for all of the potential molecular genetic abnormalities using a single test
Roberts KG et al 2016 J Clin Oncol
Ph-like lymphoblastic leukemia
Roberts et al, N Engl J Med 2014;371:1005-15
Targetable kinase activating abnormalities in Ph-like ALL
Roberts KG et al. 2014. N Engl J Med 371;11
Clinical scenario #3• 29 year old male with relapsed B-
lymphoblastic leukemia– Initial diagnosis 2012
• Negative for t(9;22);BCR-ABL1
– Tested at relapse using a single NGS panel (Foundation Medicine)
• IGH-CRLF2 rearrangement• IKZF1 deletion• PAX5 missense mutation
– Findings indicate “Ph-like” ALL– Awaiting transplant, may benefit from kinase
inhibitor therapy
Panel-based NGS testingMutation panel testing by NGS
Pros1. Variants are reported together, at the same time, on a single report2. Interpretation takes into account all variants identified3. Cost is less compared to multiple single gene tests4. Variant frequencies provide information on subclonal structure5. Pattern and identity of mutations facilitates accurate subclassification and
prognostication6. Detection of certain variants allows for the use of targeted therapies
Cons1. May not be reimbursed by payers2. Variants of unknown significance – what to do?3. Some of the information is not currently actionable
Conclusions• NGS is revolutionizing pathology and laboratory medicine
• Allows for true personalized medicine
• Facilitates use of targeted therapeutic strategies
• Costs are rapidly decreasing while the technology continues to improve
• Challenges remain– Cost and reimbursement– Data analysis – Variant interpretation – Other aspects of testing (ie. PCR) can affect the results!
• Today – panels and genetically complex single gene analysis; detection of targeted structural variants
• Future – routine comprehensive whole genome analysis of tumors